remote sensing Article Spring Frost Damage to Tea Plants Can Be Identified with Daily Minimum Air Temperatures Estimated by MODIS Land Surface Temperature Products Peijuan Wang 1,*, Yuping Ma 1 , Junxian Tang 1, Dingrong Wu 1 , Hui Chen 2, Zhifeng Jin 3 and Zhiguo Huo 1,4 1 State Key Laboratory of Severe Weather, Chinese Academy of Meteorological Sciences, Beijing 100081, China; [email protected] (Y.M.); [email protected] (J.T.); [email protected] (D.W.); [email protected] (Z.H.) 2 Fujian Institute of Meteorological Science, Fuzhou 350001, China; [email protected] 3 Zhejiang Climate Center, Hangzhou 310017, China; [email protected] 4 Collaborative Innovation Center of Meteorological Disaster Forecast, Early-Warning and Assessment, Nanjing University of Information Science & Technology, Nanjing 210044, China * Correspondence: [email protected] Abstract: Tea (Camellia sinensis) is one of the most dominant economic plants in China and plays an important role in agricultural economic benefits. Spring tea is the most popular drink due to Chinese drinking habits. Although the global temperature is generally warming, spring frost damage (SFD) to tea plants still occurs from time to time, and severely restricts the production and quality of spring tea. Therefore, monitoring and evaluating the impact of SFD to tea plants in a timely and precise manner is a significant and urgent task for scientists and tea producers in China. The region designated as Citation: Wang, P.; Ma, Y.; Tang, J.; the Middle and Lower Reaches of the Yangtze River (MLRYR) in China is a major tea plantation area Wu, D.; Chen, H.; Jin, Z.; Huo, Z. producing small tea leaves and low shrubs. This region was selected to study SFD to tea plants using Spring Frost Damage to Tea Plants meteorological observations and remotely sensed products. Comparative analysis between minimum Can Be Identified with Daily air temperature (Tmin) and two MODIS nighttime land surface temperature (LST) products at six pixel- Minimum Air Temperatures window scales was used to determine the best suitable product and spatial scale. Results showed Estimated by MODIS Land Surface that the LST nighttime product derived from MYD11A1 data at the 3 × 3 pixel window resolution Temperature Products. Remote Sens. was the best proxy for daily minimum air temperature. A Tmin estimation model was established 2021, 13, 1177. https://doi.org/ using this dataset and digital elevation model (DEM) data, employing the standard lapse rate of 10.3390/rs13061177 air temperature with elevation. Model validation with 145,210 ground-based Tmin observations showed that the accuracy of estimated Tmin was acceptable with a relatively high coefficient of Academic Editor: Nicolas R. Dalezios determination (R2 = 0.841), low root mean square error (RMSE = 2.15 ◦C) and mean absolute error ◦ Received: 1 February 2021 (MAE = 1.66 C), and reasonable normalized RMSE (NRMSE = 25.4%) and Nash–Sutcliffe model Accepted: 14 March 2021 efficiency (EF = 0.12), with significantly improved consistency of LST and Tmin estimation. Based on Published: 19 March 2021 the Tmin estimation model, three major cooling episodes recorded in the “Yearbook of Meteorological Disasters in China” in spring 2006 were accurately identified, and several highlighted regions in Publisher’s Note: MDPI stays neutral the first two cooling episodes were also precisely captured. This study confirmed that estimating with regard to jurisdictional claims in Tmin based on MYD11A1 nighttime products and DEM is a useful method for monitoring and published maps and institutional affil- evaluating SFD to tea plants in the MLRYR. Furthermore, this method precisely identified the spatial iations. characteristics and distribution of SFD and will therefore be helpful for taking effective preventative measures to mitigate the economic losses resulting from frost damage. Keywords: spring frost damage; MODIS LST product; minimum air temperature; tea plants; the Copyright: © 2021 by the authors. Middle and Lower Reaches of the Yangtze River Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons 1. Introduction Attribution (CC BY) license (https:// Tea plant (Camellia sinensis) is a species of evergreen shrubs or small trees in the creativecommons.org/licenses/by/ flowering plant family Theaceae whose leaves and leaf buds are used to produce tea 4.0/). Remote Sens. 2021, 13, 1177. https://doi.org/10.3390/rs13061177 https://www.mdpi.com/journal/remotesensing Remote Sens. 2021, 13, 1177 2 of 22 drink. Tea, along with coffee (Coffea arabica L.) and cocoa (Theobroma cacao L.), has been used to make most of the traditional and popular caffeinated beverages [1] in the world. Differences in tea taste are determined by the degree of fermentation. There are seven classifications of tea, including black tea, white tea, yellow tea, green tea, oolong tea, dark tea, and reprocessing tea, based on how they are processed and how long the leaves are left on the shrub before harvesting [2]. Tea plant originated in southeast China and was gradually introduced into many tropical and subtropical countries [3]. At present, more than 60 countries and regions around the world plant tea trees, mainly distributed in tropical and subtropical climate areas between 16◦S and 30◦N[4]. Some varieties also grow well in cooler climates and high altitude regions, such as the mild climate of the U.S. Pacific Northwest and the Yunnan–Guizhou Plateau of southwestern China. The worldwide harvested tea area reached 3.56 × 106 ha in 2018 with a production of 5.11 × 109 kg [5]. Of all of the countries of the world, China harvested the greatest area (66.0% of the world total) and produced the greatest amount (51.4% of the world total) of tea, followed by India. These two countries accounted for 83.7% of the world tea harvested area and 77.8% of the world tea production [5]. The 18 provinces in China where planting of tea plants is concentrated are located in North China, Central China, South China, and Southwest China. Tea can be divided into four categories according to the harvest season: spring tea, summer tea, autumn tea, and winter tea. Spring tea is usually harvested from March to the middle of May. Compared with teas harvested in other seasons, the spring leaves are softer, the buds are plumper, the color is greener, the vitamins and amino acids are richer, and the flavor tastes more refreshing due to moderate temperatures and higher relative air humidity in spring [6,7]. Therefore, spring tea is an important segment of the tea market because of its high quality and the highest economic benefits for tea production all over the world. For example, the economic benefits of spring tea accounted for 87.5% of the total tea revenue in 2019 in Zhejiang Province in China [8]. According to climate model projections, the global average air temperature is likely to increase by 0.3–4.8 ◦C by the end of this century [9]. However, research in recent years has also revealed that the frequency, intensity, and duration of extreme low temperature events are also likely to continuously increase under 21st-century warming scenarios [10,11]. Additionally, the predicted spring frost risk correspondingly increases under future climate warming [12]. Frost damage can be divided into advection frost and radiation frost according to its formation mechanism [13]. An advection frost occurs when cold air blows into an area to replace warmer air that was present before the weather change. It is associated with moderate to strong winds and low humidity [13]. Radiation frosts occur because of heat losses in the form of radiant energy. Under clear nighttime skies, the soil surface loses more longwave radiation (heat) to the atmosphere than it receives [14], resulting in declining temperature. Radiation frosts are characterized by clear skies, calm winds, and temperature inversions [13]. During a radiation frost event, the ground and ambient air temperatures drop, and vegetation canopies experience temperatures that are 2–4 ◦C colder than the air temperature that is measured within a nearby standard Stevenson Screen thermometer enclosure. Spring frost is a catastrophic weather event that may cover a broad spatial region and is harmful to crops (winter wheat [15], winter sugar beet [16], potato [17]); orchards (apple [18], cherry [18], tea [19]); and forests (beech [20–22]), spruce [18]). Different methods of spring frost identification and quantification have been studied in recent years, and generally fall into two categories: one is the spring frost index based on air temperature (Ta) measured with sensors inside meteorological station shelters at a height of 2 m above the ground surface, and the other is spring frost identification based on the land surface temperature (LST) derived from satellite remote sensing imagery. Ground-based meteorological observations showed that when daily minimum air temperature is less than 4 ◦C in spring, newly formed tea buds were impacted by spring frost [23], and the commod- ity value of the tea declined correspondingly. Therefore, tea leaf damage grades have been Remote Sens. 2021, 13, 1177 3 of 22 quantified based on daily minimum air temperature and its duration [24]. Li et al. [25] and Jin et al. [26] further refined this frost damage indicator by using hourly temperature and its duration. Meteorological data generally has high accuracy, high temporal resolution, and long-term observation datasets, and has therefore been used in many applications with regards to spring frost, such as identifying spatiotemporal distribution characteristics [27], and assessing impact [28] and risk [29]. Nevertheless, observation of Ta is limited by weather station network layout, resulting in point samples that often cannot effectively reflect spatial variation, especially in certain areas where the distribution of meteorological stations is itself very sparse [30], and in areas where elevation changes drastically.